Measuring problems in estimating the exposure to passive smoking using the excretion of cotinine.

Quality control studies on cotinine measurements following low level environmental tobacco smoke (ETS) exposure are rare. The exposure to ETS was controlled and systematically changed in a series of experiments in a climatic chamber. Healthy nonsmoking volunteers were exposed to ETS simultaneously. The duration and level of exposure varied using high (8, 17 and 25 ppm CO), and low (2 and 5 ppm CO) exposure levels. The variation between radioimmunoassay (RIA) and gas chromatography (GC) was high as was the variation between the results of RIA laboratories. There was also a high within-laboratory-variation. A 1:10 dilution seems to be preferable over a 1:3 dilution. Freezing the urine samples immediately after collection led to the detection of higher cotinine values than freezing the samples 24 h after collection. Highly reliable data for cotinine were obtained when the urine samples were kept frozen immediately after collection and fractionated sampling over 48-72 h was used. Our data show that estimating low-level ETS exposure by measuring urinary cotinine is highly susceptible to uncontrolled variation and errors. Sufficiently reliable estimates of low-level ETS exposure can be made only when fractionated sampling over 48-72 h is used and when the urine samples are kept frozen just after collection.